Effects of clinical pathways in stroke management: A meta-analysis

Neurology Asia 2015; 20(4) : 335 342 Effects of clinical pathways in stroke management: A meta-analysis 1,2,3 Di Huang, 1,2,3 XuPing Song, 2,3 Jinhui Tian, 1,2,3 Qi Cui, 1,2,3 Kehu Yang 1 School of Public Health, Lanzhou University, Lanzhou, Gansu Province; 2 Evidence-Based Medicine Center, School of Basic Medical Sciences, Lanzhou University, Lanzhou, Gansu Province, China; 3 Key Laboratory of Evidence Based Medicine and Knowledge Translation of Gansu Province, Lanzhou, Gansu Province, China Abstract Objective: To assess the implementation effects of clinical pathways, compared with usual care, among patients with stroke. Methods:Two investigators independently searched PubMed, Embase, the Cochrane Library, Web of Science, Chinese Biomedical Literature Database and Wanfang Database for studies published before December 2014. Jadad methodological approach was applied to assess the quality of included studies and RevMan software (version 5.2.7) was used for meta-analysis. Results: A total of 11 RCTs involving 913 patients were included in this meta-analysis. The overall results showed that a shorter average length of stay [MD = -2.92; 95% CI (-4.06, -1.78); P < 0.001] and a lower inpatient expenditures [SMD = -1.64; 95% CI (-1.80, -1.48); P < 0.001] in clinical pathways group comparing with the usual care group. The higher score of patient satisfaction was also seen in clinical pathways group. Conclusion: clinical pathways may reduce the average length of stay, reduce the inpatient expenditures, increase patient satisfaction and improve the quality of care in stroke management. INTRODUCTION Stroke is a global epidemic and a major public health-care concern as mortality rates are known to vary greatly between countries and geographic regions. Annually, 15 million people worldwide suffer an episode of stroke. Of these, 5 million die and another 5 million are left permanently disabled, with added burden to the family and community. 1 Clinical pathways (CPW) originated in the United States in the 1980s; these were developed through the collaborative efforts of physicians, nurses, pharmacists, and others to improve the quality and value of patient care. CPW map a patient s journey, providing coordination of services for users and aim to have the right people, doing the right things, in the right order, at the right time, in the right place, with the right outcome. 2 It is a tool to provide standard care with little variability. 3 Several researchers 4-7 have shown that the implementation of clinical pathways reduce the variability of clinical practice and improve outcomes. In the United States. CPW have been applied to health care management since the 1980s to improve efficiency of care and reduce hospitalization costs. CPW is increasingly being implemented in many countries to improve the care of stroke patients, but the effects of CPW on stroke management are not clear. This meta-analysis aims to assess the effects of CPW on stroke management in the hospital, and to evaluate whether CPW improve the outcomes of patients who have a stroke. METHODS We used the methods described in the preferred reporting items for systematic reviews and metaanalyses (PRISMA) statement. 8 Literature search strategy Two investigators independently searched PubMed, Embase, the Cochrane Library, Web of Science, Chinese Biomedical Literature Database, and Wanfang Database for studies published before December 2014. We searched the electronic databases combining medical subject headings (MeSH) terms with free text terms. MeSH terms were performed based on the following search string: clinical pathways, critical pathways, care pathways, care map, and stroke, Address correspondence to:kehu Yang, Professor of Medicine, Chief, Evidence-Based Medicine Center, School of Basic Medical Sciences, Lanzhou University, Lanzhou730000, Gansu Province, China. Tel: +86-18394170221, e-mail: kehuyangebm2006@126.com 335

Neurology Asia December 2015 cerebral vascular accident, apoplexy, and randomized. Details of the search strategy used in PubMed are shown in the Appendix. There were no language restrictions during the document retrieval. In addition, the reference list of the retrieved articles was also searched and relevant studies were checked manually to identify other studies pertaining to the study topic. Selection criteria We identified and reviewed all studies that met the following criteria: (1) Type of study: prospective randomized controlled trials, (2) Participants: all studies that recruited patients who had been admitted to hospital with a clinical diagnosis of stroke, (3) Intervention: patients who were diagnosed as having astroke were randomized either to the clinical pathways (CPW) group or to the non-intervention (usual care) group; (4) Outcome measures: average hospital length of stay, inpatient expenditures, and patient satisfaction. Abstracts, case reports, conference presentations, editorials, and expert opinions were excluded. Data extraction and critical appraisal Two reviewers (Huang D and Song XP) independently screened the article titles, abstracts, and full texts to identify studies that met our inclusion criteria. Any conflicts or disagreements were adjudicated by a third reviewer (Tian JH). The following variables were extracted from all studies: first author, publication year, study design, country, participant information (sample size, mean age, percentage of men, and baseline information), and reported indicators. We contacted authors for additional data or clarification where needed. Extracted data were entered into Microsoft Office Excel 2007 (Microsoft, Seattle WA, USA) and were checked by the third reviewer (Tian JH). Discrepancies between reviewers assessments of the publications, conceptual problems on the pathway intervention or methodological and statistical problems were solved through discussion, or advice was sought from a third reviewer (Tian JH). Statistical analysis This meta-analysis was performed using RevMan 5.2.7 software (available from the website for free: http://ims.cochrane.org/revman). In the meta-analysis, the Mantel-Haenszel method was used for analyses of dichotomous data and results were presented as Odds Ratios (ORs) with 95% confidence intervals (CI). The inverse variance method was used to pool continuous data, and the mean difference (MD) with 95% confidence interval (CI) was used to evaluate continuous data when all studies use the same scale to report their outcomes, while the standardized mean difference (SMD) with 95% CI was used to evaluate the studies using different scales. The heterogeneity between trials was assessed by the Chi-square test and the extent of inconsistency was evaluated by the I 2 statistic. Meta-analysis was performed using a fixed-effects model (p>0.1) or a randomizedeffects model (p 0.1) according to the degree of heterogeneity. Sensitivity analyses were carried out to investigate the influence of each study on the overall outcome of the meta-analysis, which were calculated using the software Meta-analyst 3.13. RESULTS Quantity and quality of trials A systematic review of the six electronic database searches identified 913 potentially relevant references. After exclusion of duplicate or irrelevant references, 267 potentially relevant articles were retrieved for more detailed evaluation. After applying the selection criteria, 11 comparative studies remained eligible for quantitative assessment. A PRISMA chart summarizing the search strategy is presented in Figure 1 and characteristics of included trials are summarized in Table 1. The risk of bias was independently assessed by two reviewers (Huang D and Song XP). The evaluation of risk of bias in included studies was assessed by the modified Jadad 7-point scale. 9 The Jadad approach is a five-point scale that assigns points to each study on the basis of the quality of the randomization generation (0 2 points), of the blinding process (0 2 points), of the description of withdrawals and dropouts (0 1 point), and of the allocation concealment (0 2 points). In general, a total score of 4 or more points was achieved only by high quality studies. Any disagreements were resolved in consultation with the third reviewer (Tian JH) who acted as an arbiter. Characteristics of included studies Characteristics of included trials are summarized in Table 1. These 11 RCTs 6,10-19 recruited a total of 984 patients. Eight studies had recruited their 336

Figure 1. Flow chart of the study selection process. Table 1: Characteristics of the included studies Study Age (years) Sample Size Staff received Year name usual training CPW usual care CPW care Reported indicators Panella 7 2011 74.5±10.8 74.0 ± 11.7 238 238 recevied ALOS, Mortality, Complication rates, Lu 11 2013 55±72 56±74 40 40 recevied ALOS, IE, PS, Sun 12 2013 47±70 48±71 25 25 recevied ALOS, IE, PS, Li 13 2011 42.3±6.8 42.6±7.2 30 30 recevied ALOS, IE, PS, GHK Bai 14 2011 59.9 * 61.2 * 100 100 Not reported ALOS, IE, PS, FAM, BI Qian 15 2011 65.63±11.93 68.72±12.98 45 45 Not reported ALOS, IE, NIHSS Wang 16 2010 50 * 50 * 43 40 Not reported ALOS, IE, NIHSS Guo 17 2010 63.5±9.2 63.5±9.2 40 40 Not reported ALOS, IE, PS, SAS He 18 2008 58.5 * 58.5 * 30 30 Not reported ALOS, IE, PS, SAS Wang 19 2007 61.3±13.2 64.8±11.7 51 50 Not reported ALOS, IE, PS, GHK Chen 20 2006 Not reported Not reported 90 90 Not reported ALOS, IE, PS, GHK *: mean; IE: Inpatient expenditures; ALOS: Average length of stay; GHK: Grasp of health knowledge; PS: Patient satisfaction; NIHSS: the National Institutes of Health Stroke Scale; BI: Barthel index; SAS: Self-rating anxiety scale; FAM: Fugl Meyer assessment 337

Neurology Asia December 2015 Table 2: Methodological quality of included studies Study name Randomization Blinding Withdrawals and dropouts Allocation concealment Total Panella 6 Not reported Not reported No Not reported 4 Lu 10 Not reported Not reported No Not reported 3 Sun 11 Not reported Not reported No Not reported 3 Li 12 Not reported Not reported No Not reported 3 Bai 13 Not reported Not reported No Not reported 3 Qian 14 Random number table Not reported No Not reported 4 Wang 15 Not reported Not reported No Not reported 3 Guo 16 Not reported Not reported No Not reported 3 He 17 Not reported Not reported No Not reported 3 Wang 18 Not reported Not reported No Not reported 3 Chen 19 Not reported Not reported No Not reported 3 participants in 2010 and later 6,10-16 and three studies recruited before 2010. 17-19 Ten studies 10-19 took place in China and one 6 in Italy. Participants were between 18 to 86 years of age. Ten studies 10-19 diagnosed a stroke based on the head computed tomography scan (CT) or magnetic resolution imaging (MRI). Nine studies 10,11,13-19 showed no significant statistical difference between CPW and usual standard of care on baseline data of patient characteristics. Table 1 shows the characteristics of the studies that were included. The evaluation of the risk of bias in the studies included in this meta-analysis is showed in Table 2. Average length of stay Ten studies 6,10-12,14-19 reported the average length of stay (LOS), of which nine studies 6,10-12,14-16,18,19 reported this indicator with mean±standard deviation (SD), the aggregate results showed that significant heterogeneity existed in included studies (I 2 = 89%; P < 0.001). Data from these studies were pooled using a random effects model; the pooled analysis indicated that the mean length of hospital stay was significantly shorter in the CPW group, [MD= -2.92; 95% CI (-4.06, -1.78); P < 0.001](Figure2). The sensitivity analysis (leaving one out at a time) produced no statistically significantly increased or decreased summary results, although the findings from the study 6 were quite influential (Figure 3). Only one study 17 did not report standard deviations and therefore could not be included for summary analysis. It showed a shorter mean length of hospital stay in the CPW group (the mean LOS of CPW group was 28 days vs. 34 days for the usual care group, P< 0.01). CPW Usual care Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Random, 95% CI IV, Random, 95% CI Chen 2006 16.3 2.3 90 20.5 3.4 90 12.3% 4.20 [ 5.05, 3.35] Guo 2010 15.61 1.6 40 20.14 3.15 40 11.8% 4.53 [ 5.62, -3.44] Li 2011 7.01 1.06 30 11.15 2.02 30 12.4% 4.14 [-4.96, 3.32] Lu 2013 11.4 1.9 40 14.5 3.3 40 11.6% 3.10 [ 4.28, 1.92] Panella 2012 11.78 6.6 238 10.88 7.9 238 11.3% 0.90 [ 0.41, 2.21] Qian 2011 14.32 2.89 45 16.75 3.25 45 11.4% 2.43 [ 3.70, 1.16] Sun 2013 13.4 1.9 25 16.5 3.3 25 10.8% 3.10 [ 4.59, 1.61] Wang 2011 16.3 8.1 43 21.4 8.3 40 5.8% 5.10 [ 8.63, 1.57] Wang 2007 18.6 1.58 51 20.1 2.09 50 12.6% 1.50 [ 2.22, 0.78] Total (95% CI) 602 598 100.0% 2.92 [ 4.06, 1.78] Heterogeneity: Tau 2 = 2.56; Chi 2 = 75.38, df = 8 (P < 0.00001); I 2 = 89% Test for overall effect: Z = 5.02 (P < 0.00001) Figure 2. Forest plot of meta-analysis: ALOS (Average Length of Stay), MD (mean difference) with 95% CI. 338

Study Excluded Chen (2006) Guo (2010) Li (2013) Lu (2013) Panella (2012) Qian (2011) Sun (2013) Wang (2011) Wang (2007) Overall N 180 80 60 80 476 90 50 83 101 Confidence Interval 2.746 ( 4.002, 1.489) 2.707 ( 3.927, 1.487) 2.754 ( 4.020, 1.489) 2.904 ( 4.190, 1.619) 3.378 ( 4.313, 2.444) 2.989 ( 4.260, 1.719) 2.904 ( 4.160, 1.648) 2.787 ( 3.966, 1.607) 3.125 (-4.325, 1.925) 2.922 ( 4.063, 1.781) Figure 3. The results of sensitivity analysis. 4 3 2 1 0 1 Inpatient expenditures Eight studies 10-14,17-19 reported inpatient expenditures. There was significant heterogeneity in included studies (I 2 = 89%; P < 0.001). Data from these studies were pooled with a randomeffects model. The aggregate results of randomeffects model showed that CPW was superior to usual care on inpatient expenditures [SMD = -1.82; 95% CI (-2.33, -1.30) (Figure 4). Patient satisfaction Patient satisfaction was reported by eight studies, five 10,11,13,16,17 of which reported in percentage and three 12,18,19 reported in mean ± standard deviation (SD). The aggregate results of five studies 10,11,13,16,17 reported in percentage presented higher patient satisfaction in CPW [OR = 2.04; 95% CI (1.04, 3.98); P=0.04] (Figure 5). The three 12,18,19 RCTs reported in mean ± SD showed CPW Usual care Std. Mean Difference Std. Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI Bai 2011 5,963.3 495.6 100 6,969.5 554.2 100 23.5% 1.91 [ 2.24, 1.57] Chen 2006 2,790.5 158.65 90 3,042.4 160.58 90 23.5% 1.57 [ 1.91, 1.24] He 2008 8,500 800 30 9,300 1,000 30 9.3% 0.87 [ 1.40, 0.34] Li 2011 3,541.05 691.07 30 4,157.93 765.82 30 9.4% 0.83 [ 1.36, 0.31] Lu 2013 7,682 375.7 40 9,725 750.3 40 5.4% 3.41 [ 4.11, 2.71] Qian 2011 8,233.71 1,684.71 45 10,732.32 2,146.89 45 12.7% 1.28 [ 1.74, 0.83] Sun 2013 7,648 358.7 25 9,873 758.3 25 3.0% 3.69 [ 4.63, 2.76] Wang 2007 2,790.5 158.65 51 3,042.4 160.58 50 13.1% 1.57 [ 2.01, 1.12] Total (95% CI) 411 410 100.0% 1.64 [ 1.80, 1.48] Heterogeneity: Chi 2 = 65.22, df = 7 (P < 0.00001); I 2 = 89% Test for overall effect: Z = 19.81 (P < 0.00001) Figure 4. Forest plot of meta-analysis: inpatient expenditures, MD (mean difference) with 95% CI. 339

Neurology Asia December 2015 CPW Usual care Odds Ratio Odds Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Bai 2011 98 100 94 100 15.0% 3.13 [0.62, 15.89] Guo 2010 35 40 32 40 32.0% 1.75 [0.52, 5.90] He 2008 25 30 24 30 32.0% 1.25 [0.34, 4.64] Lu 2013 38 40 35 40 14.0% 2.71 [0.49, 14.90] Sun 2013 24 25 22 25 7.0% 3.27 [0.32, 33.84] Total (95% CI) 235 235 100.0% 2.04 [1.04, 3.98] Total events 220 207 Heterogeneity: Chi 2 = 1.13, df = 4 (P = 0.89); I 2 = 0% Test for overall effect: Z = 2.09 (P = 0.04) Figure 5. Forest plot of meta-analysis: patient satisfaction (%), OR with 95% CI. CPW Usual care Std. Mean Difference Std. Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI Chen 2006 29.6 5.35 90 21 4.98 90 51.7% 1.66 [1.32, 200] Li 2011 98.24 1.65 30 95.16 3.12 30 19.4% 1.22 [0.66, 1.77] Wang 2007 29.6 5.35 51 21 4.98 50 28.9% 1.65 [1.20, 2.10] Total (95% CI) 171 170 100.0% 1.57 [1.33, 1.81] Heterogeneity: Chi 2 = 1.92, df = 2 (P = 0.38); I 2 = 0% Test for overall effect: Z = 12.61 (P < 0.00001) Figure 6. Forest plot of meta-analysis: patient satisfaction (Mean±SD), standardized mean difference (SMD) with 95% CI. that patients were significantly more satisfied with their hospital care in the CPW group [SMD = 1.57; 95% CI (1.33, 1.81); P<0.0001] (Figure 6). DISCUSSION CPW was initially employed in 1980s in the United States. They were designed to minimize delays and resource utilization, maximize quality of care, reduce variation in care and provide continuous quality improvement. 20 The results of our meta-analysis show that CPW was effective in improving the outcomes of patients who had a stroke. CPW was associated with a significantly decreased average length of stay and reduction in inpatient expenditures compared to the usual care. One of the studies included in our analysis 6 showed that usual care was superior to CPW in the average length of stay, but the difference did not achieve statistical significance. In addition, the sensitivity analysis showed that our result was stable. The average length of stay and inpatient expenditures were measured as days and yuan (the monetary unit of China), which were the most commonly employed outcomes and from the economic perspective. As a major indicator reflecting quality of care, evidence from eight randomized studies 10-13,16-19 reported the patient s satisfaction. The aggregate results showed that CPW had a positive effect on patient satisfaction compared with the usual care. Only one study 18 reported patient satisfaction quantitatively, with a total score 30 points, while others did not mention any quantitative score. In addition, four studies 10,11,18,19 measured this indicator by a comprehensive scale and others did not describe the questionnaires. Future research should report more quantitative details on this aspect of study. The strength of our study is the large number of patients. In addition, many previous systematic reviews focused on investigating CPW for stroke rehabilitation or the shoulder pain after stroke 21,22 ; whereas our paper aimed to study the effects of CPW in stroke management. However, the limitations in the present metaanalysis should be acknowledged. Some factors added to the difficulty in interpreting the results of this review and high heterogeneity existed in several pooled results. First, random sequence generation was described in only one study. 14 Second, the studies included in our meta-analysis were from different countries and different hospitals, which may be the potential sources of heterogeneity. Third, the development of pathways has been described in great detail, but little information was provided on the logistics of their implementation; this may have affected the outcomes. Moreover, publication bias may have influenced the results and outcomes of the studies, as authors may prefer to write no difference 340

Appendix 1: Search strategy for PubMed Search Queries #1 critical pathways [MeSH] #2 (clinical pathway OR clinical pathways OR clinical path OR clinical paths) [Tiltle/Abstract] #3 (critical pathway OR critical pathways OR critical path OR critical paths) [Tiltle/Abstract] #4 (care pathway OR care pathways OR care path OR care paths) [Tiltle/Abstract] #5 (care map OR care maps) [Tiltle/Abstract] #6 (care protocol OR care protocols) [Tiltle/Abstract] #7 #1 OR #2 OR #3 OR #4 OR #5 OR #6 #8 Stroke [MeSH] #9 apoplexy [MeSH] #10 (stroke* OR cerebrovascular* OR cerebral vascular OR cva*) [Tiltle/Abstract] #11 (cerebral OR cerebellar OR brainstem OR vertebrobasilar ) [Tiltle/Abstract] #12 (infarct* OR isch?emi* OR thrombo* OR apoplexy OR emboli*) [Tiltle/Abstract] #13 #11 AND #12 #14 (cerebral OR intracerebral OR intracranial OR parenchymal) [Tiltle/Abstract] #15 (brain or intraventricular or brainstem or cerebellar) [Tiltle/Abstract] #16 (infratentorial or supratentorial or subarachnoid) [Tiltle/Abstract] #17 #14 OR #15 OR #16 #18 (haemorrhage or hemorrhage or haematoma or hematoma) [Tiltle/Abstract] #19 (bleeding or aneurysm) [Tiltle/Abstract] #20 #18 OR #19 #21 #17 AND #20 #22 #8 OR #9 OR #10 OR #13 OR #21 #23 #7 AND #22 rather than report the actual data or they may omit publication of studies with negative results. Finally, methods were not clearly described in some of the studies and this may resulted in confounding factors, and sources of contamination that have not been identified. In conclusion, the results of this meta-analysis demonstrate that CPW improves the quality of care in patients with stroke. Theoretical advantages of such methods are clear. However the benefits may be less than expected because of variations in patients characteristics, pre-existing practices, and/or dependence on external factors. Further RCTs are needed before implementation of the CPW in stroke patients. We should also pay attention to the following aspects: the authors of CPW should receive training in quality improvement and in the development of CPW before implementation. Establishing an evidence based CPW should be a focus for the future. Moreover, there is currently limited research on the formulation of CPW; additional studies may provide evidence based medicine and a beneficial direction for formulating CPW. REFERENCES 1. Shrivastava SR, Shrivastava PS, Ramasamy JD. Reduction in global burden of stroke in underserved areas. J Neurosci Rural Pract 2013; 4:475-76. 2. Allen D, Rixson L. How has the impact of care pathway technologies on service integration in stroke care been measured and what is the strength of the evidence to support their effectiveness in this respect? Int J Evid Based Healthc 2008; 6:78-110. 3. Martínez-Sánchez P, Fuentes B, Medina-Báez J, et al. Development of an acute stroke care pathway in a hospital with stroke unit. Neurología2010; 25:17-26. 4. Preston SR, Markar SR, Baker CR, et al. Impact of a multidisciplinary standardized clinical pathway on perioperative outcomes in patients with oesophageal cancer. Br JSurg 2013; 100:105-12. 5. Seo HS, Song KY, Jeon HM,et al. The impact of an increased application of critical pathway for gastrectomy on the length of stay and cost. J Gastric Cancer 2012; 12:126-31. 6. Panella M, Marchisio S, Brambilla R, et al. A cluster randomized trial to assess the effect of clinical pathways for patients with stroke: results of the clinical pathways for effective and appropriate care study. BMC Med 2012; 10:71. 7. Song PP, Gao JJ, Kokudo N, et al. Knowledge into action Exploration of an appropriate approach for constructing evidence-based clinical practice guidelines for hepatocellular carcinoma.biosci Trends 2012; 6:147-52. 341

Neurology Asia December 2015 8. Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for systematic reviews and metaanalyses: the PRISMA statement.int J Surg 2010; 8:336-41. 9. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary?control Clin Trials 1996; 17:1-12. 10. Lu H. Analysis of clinical nursing path effect in patients with cerebral infarction. Health Must-read 2013; 12:484-85. 11. Sun XJ. Application effectiveness analysis in clinical pathways in patients with cerebral hemorrhage. Health Required 2013; 12:192-93. 12. Li PZ. Observation of the effect of Clinical pathways used in transient ischemic attack. Chinese Manipulation Rehabilitation Medicine2011; 2:43-4. 13. Bai WT, Yang LR. The implementation of the effect of clinical pathway for the treatment of hypertensive cerebral hemorrhage. Chinese Community Doctors 2011; 13:31-2. 14. Qian J. To analysis the effects of a clinical pathway for cerebral infarction in participation of the clinical pharmacists.practical Clinical Medicine 2011; 12:7-11. 15. Wang WW. The clinical effect of cerebral infarction patients with the clinical pathway management analysis.china Foreign Medical Treatment 2010; 29:96-7. 16. Guo XQ. Clinical pathway in patients with cerebral infarction.chinese J Trauma Disability Medicine 2010; 18:101-02. 17. He Y. Clinical pathway in patients with cerebral hemorrhage. J Medical Theory Practice 2008; 21:473 74. 18. Wang YQ, Chu PH, He R. Clinical pathways in hemiplegic stroke prevention in patients with shoulder pain. Int J Nursing 2007; 26:1174-176. 19. Chen SL. The clinical effect in patients with cerebral infarction path analysis. Clin J Med Offic2006; 34:360-61. 20. Burns LR, Denton M, Goldfein S, Warrick L, Morenz B, Sales B. The use of continuous quality improvement methods in the development and dissemination of medical practice guidelines. QRB Qual Rev Bull 1992; 18:434-39. 21. Kwan J. Care pathways for acute stroke care and stroke rehabilitation: from theory to evidence. J ClinNeurosci 2007; 14:189-200. 22. Turner-Stokes L, Jackson D. Shoulder pain after stroke: a review of the evidence base to inform the development of an integrated care pathway. ClinRehabil 2002; 16: 276-98. 342